Skip to main content

Full text of "Diet, feeding behavior, growth, and numbers of a population of Cerberus rynchops (Serpentes: Homalopsinae) in Malaysia"

See other formats


UNIVERSITY OF 
•I l INOIS LIBRARY 

ft ' BIOLOGY 

APR 91992 



FIELD 




Zool 



NEW SERIES, NO. 50 




Diet, Feeding Behavior, Growth, and Numbers 
of a Population of Cerberus rynchops 
(Serpentes: Homalopsinae) in Malaysia 



Bruce C. Jayne 
Harold K. Voris 
Kiew Bong Heang 



lift Mttfte. m 



A Contribution in Celebration 

of the Distinguished Scholarship of Robert F. Inger 

on the Occasion of His Sixty-Fifth Birthday 



September 30, 1988 
Publication 1394 



PUBLISHED BY FIELD MUSEUM OF NATURAL HISTORY 



Information for Contributors to Fieldiana 

General: Fieldiana is primarily a journal for Field Museum staff members and research associates, althoi 
manuscripts from nonaffiliated authors may be considered as space permits. The Journal carries a page charge 
$65 per printed page or fraction thereof. Contributions from staff, research associates, and invited authors will be 
considered for publication regardless of ability to pay page charges, but the full charge is mandatory for nonaffiliated 
authors of unsolicited manuscripts. Payment of at least 50% of page charges qualifies a paper for expedited processing, 
which reduces the publication time. 

Manuscripts should be submitted to Dr. James S. Ashe, Scientific Editor, Fieldiana, Field Museum of Natural 
History, Chicago, Illinois 60605-2496, USA. Three complete copies of the text (including title page and abstract) 
and of the illustrations should be submitted (one original copy plus two review copies which may be machine copies). 
No manuscripts will be considered for publication or submitted to reviewers before all materials are complete and 
in the hands of the Scientific Editor. 

Text: Manuscripts must be typewritten double-spaced on standard- weight, 8 'A- by 1 1 -inch paper with wide margins 
on all four sides. For papers longer than 100 manuscript pages, authors are requested to submit a "Table of Contents," 
a "List of Illustrations," and a "List of Tables." In most cases, the text should be preceded by an "Abstract" and 
should conclude with "Acknowledgments" (if any) and "Literature Cited." All measurements should be in the metric 
system. The format and style of headings should follow those of recent issues of Fieldiana. For more detailed style 
information, see The Chicago Manual of Style (13th ed.), published by The University of Chicago Press, and also 
recent issues of Fieldiana. 

In "Literature Cited," authors are encouraged to give journal and book titles in full. Where abbreviations are 
desirable (e.g., in citation of synonymies), authors consistently should follow Botanico-Periodicum-Huntianum and 
TL-2 Taxonomic Literature by F. A. Stafleu & R. S. Cowan (1976 et seq.) (botanical papers) or Serial Sources for 
the Biosis Data Base (1983) published by the BioSciences Information Service. 
References should be typed in the following form: 
Croat, T. B. 1978. Flora of Barro Colorado Island. Stanford University Press, Stanford, Calif, 943 pp. 
Grubb, P. J., J. R. Lloyd, and T. D. Pennington. 1963. A comparison of montane and lowland rain forest 

in Ecuador. I. The forest structure, physiognomy, and floristics. Journal of Ecology, 51: 567-601. 
Langdon, E. J. M. 1979. Yage among the Siona: Cultural patterns in visions, pp. 63-80. In Browman, D. L., 

and R. A. Schwarz, eds., Spirits, Shamans, and Stars. Mouton Publishers, The Hague, Netherlands. 
Murra, J. 1946. The historic tribes of Ecuador, pp. 785-821. In Steward, J. H., ed., Handbook of South 
American Indians. Vol. 2, The Andean Civilizations. Bulletin 1 43, Bureau of American Ethnology, Smithsonian 
Institution, Washington, D.C. 

Stolze, R. G. 1981. Ferns and fern allies of Guatemala. Part II. Polypodiaceae. Fieldiana: Botanv, n.s., 6: 1- 

522. 

Illustrations: Illustrations are referred to in the text as "figures" (not as "plates"). Figures must be accompanied 
by some indication of scale, normally a reference bar. Statements in figure captions alone, such as "x 0.8," are not 
acceptable. Captions should be typed double-spaced and consecutively. See recent issues of Fieldiana for details of 
style. 

Figures as submitted should, whenever practicable, be 8V2 by 1 1 inches (22 x 28 cm) and may not exceed 1 1 
I6V2 inches (30 x 42 cm). Illustrations should be mounted on boards in the arrangement you wish to obtain in the 
printed work. This original set should be suitable for transmission to the printer as follows: Pen and ink drawings 
may be originals (preferred) or photostats; shaded drawings should be originals, but within the size limitation; and 
photostats should be high-quality, glossy, black and white prints. All illustrations should be marked on the reverse 
with author's name, figure number(s), and "top." Original illustrations will be returned to the author upon publication 
unless otherwise specified. Authors who wish to publish figures that require costly special paper or color reproduction 
must make prior arrangements with the Scientific Editor. 

Page Proofs: Fieldiana employs a two-step correction system. Each author will normally receive a copy of the 
edited manuscript on which deletions, additions, and changes can be made and queries answered. Only one set of 
page proofs will be sent. All desired corrections of type must be made on the single set of page proofs. Changes in 
page proofs (as opposed to corrections) are very expensive. Author-generated changes in page proofs can only be 
made if the author agrees in advance to pay for them. 



FIELDIANA 



Zoology 

STEW SERIES, NO. 50 



)iet, Feeding Behavior, Growth, and Numbers 
rf a Population of Cerberus rynchops 
^Serpentes: Homalopsinae) in Malaysia 

Jruce C. Jayne Harold K. Voris Kiew Bong Heang 

department of Developmental Department of Zoology Department of Biology 

and Cell Biology Field Museum of Natural History University of Malaysia 

Jniversity of California Chicago, Illinois 60605-2496 Kuala Lumpur, Malaysia 
mine, California 92717 



i Contribution in Celebration 

)f the Distinguished Scholarship of Robert F. Inger 

m the Occasion of His Sixty-Fifth Birthday 



Accepted for publication February 23, 1987 
September 30, 1988 
Publication 1394 



p 



UBLISHED BY FIELD MUSEUM OF NATURAL HISTORY 



© 1988 Field Museum of Natural History 

ISSN 0015-0754 

PRINTED IN THE UNITED STATES OF AMERICA 



Table of Contents 

;Abstract 1 

Introduction 1 

Materials and Methods 2 

Results 

j Diet 4 

Predator/Prey Size Relationships 5 

I Foraging 6 

I Feeding Behavior 7 

,! Growth 10 

ii Population Structure 11 

Population Size Estimates 11 

(Discussion 11 

Feeding Behavior 12 

Growth 13 

Reproduction and Population Numbers .13 
Vcknowledgments 14 

TERATURE ClTED 14 



Jst of Illustrations 



Schematic map of the study site in the 
Muar estuary, Maylasia 



2. Plot of log total mass of stomach con- 
tents versus log mass of snake for the 181 
Cerberus rynchops with prey items 7 

3. Striking behavior of Cerberus rynchops 
attacking a goldfish 8 



List of Tables 



Size distribution for samples of Cerberus 

rynchops 5 

Diet of Cerberus rynchops 6 

Coefficients for multiple regression equa- 
tions predicting prey handling times for 
Cerberus rynchops consuming Perioph- 

thalmus chrysospilos 9 

Predicted handling times for Cerberus 
rynchops eating Periophthalmus chrysos- 
pilos 10 

Snout- vent lengths and masses for 14 of 
the 24 recaptured Cerberus rynchops with 
the greatest percentage increase in mass . 10 



4. 



in 



Diet, Feeding Behavior, Growth, and Numbers 
of a Population of Cerberus rynchops 
(Serpentes: Homalopsinae) in Malaysia 



Abstract 

Stomach contents were obtained from 181 of 
6 1 1 Cerberus rynchops captured near the mouth 
of the Muar River in Malaysia. Of the prey items, 
69% were the goby Oxuderces dentatus; however, 
as C rynchops become larger, ariid catfish, mullet, 
and taenioid gobies are increasingly important 
portions of the diet. These species of prey, com- 
bined with direct observations, suggest that C. ryn- 
chops usually forage on or near the bottom or in 
very shallow water. Feeding behavior was ob- 
served for 23 C rynchops which consumed 71 
mudskippers. Initial seizure of the fish always in- 
volved marked lateral flexion of the neck. Snakes 
often held the fish before the initiation of swal- 
lowing, and regression analysis revealed holding 
was significantly longer with prey of larger size, 
which struggled more. The venom apparatus of C. 
rynchops is capable of immobilizing and killing 
fish smaller than 3 g. Recapture of 24 tagged snakes 
allowed estimation of average percentage of growth 
rates in snout-vent length (x = 0.09%/day, range 
0%-0.27%/day) and mass (x = 0.36%/day, range 
-0.28%-1.43%/day). No evidence of a seasonal 
reproductive pattern was found. These aspects of 
the natural history of C. rynchops are compared 
to those of the sympatric species of marine snakes. 



Introduction 

The diverse assemblage of southeast Asian ma- 
rine snakes includes three distinct taxonomic 
groups; homalopsines, acrochordids, and hydro- 
phiids. Cerberus rynchops is one of several hom- 
alopsine species that is abundant in a variety of 



coastal habitats (Wall, 1918; Smith, 1943; Gyi, 
1 970; Tweedie, 1 983). Dunson and Minton (1978) 
found C. rynchops co-occurring with acrochordids 
and hydrophiids in mangrove areas of the Phil- 
ippines. Cerberus rynchops can excrete salt via a 
salt gland (Dunson & Dunson, 1979) and can ac- 
quire some oxygen through cutaneous uptake 
(Heatwole & Seymour, 1978); however, these ca- 
pacities apparently are not as well developed as 
those of the hydrophiids. Smith (1943) reported 
that C. rynchops is piscivorous, but the lack of 
data on the species of prey prevents determination 
of dietary overlap with sympatric species of ma- 
rine snakes. Cerberus rynchops possesses opistho- 
glyphous dentition, but the role of this dentition 
in prey capture and manipulation is not known. 
The biological role of opisthoglyphous dentition 
is of considerable interest when discussing the evo- 
lution of ophidian venom apparatus (Kardong, 
1 980). Cerberus rynchops also has a relatively stout 
body and fragmented head scalation (Gyi, 1970), 
characters that Pough and Groves (1983) corre- 
lated with the proficient handling of large prey by 
snakes. However, the size of natural prey of C. 
rynchops has not been documented. 

Most information on the natural history of ma- 
rine snakes pertains to the hydrophiids. Dunson 
(1975) recently reviewed much of the literature 
and provided most information for Australian 
species. Much information about southeast Asian 
species has come from studies of the hydrophiids 
occurring in the main channel of the mouth of the 
Muar River in Johore, Malaysia. These reports 
have investigated diet (Voris & Voris, 1 983), feed- 
ing behavior (Voris et al., 1978), reproduction, 
growth, and population size (Voris & Jayne, 1 979; 
Lemen & Voris, 1981; Voris, 1985). 

The purpose of this study is to expand the 



JAYNE ET AL.: CERBERUS RYNCHOPS 



knowledge of the marine snake fauna of the Muar 
River estuary by investigating several aspects of 
the natural history of Cerberus rynchops which 
occurs on the adjacent intertidal mud flat. First, 
the species composition and size of prey items are 
determined. Second, feeding behavior is analyzed, 
particularly to determine the role of opisthoglyph- 
ous venom apparatus and to quantify the effect of 
prey size on handling time. Third, growth rates 
and population size are estimated. Finally, com- 
parisons are made with the sympatric species of 
marine snakes. 



Materials and Methods 

The fieldwork for this project extended from 1 4 
January-5 March 1984 and 20 November 1985— 
8 February 1986. The primary site for this work 
was 400 m of shoreline along the south side of the 
mouth of the Muar River, Johore, Malaysia (fig. 
1). Habitats at the shoreline include mud, sand, 
and gravel beach, man-made stone walls, and dense 
mangrove. From these varied shore habitats, the 
intertidal zone extends seaward as one continuous 
mud flat about 50 m wide at the mouth of the 
river to over 200 m wide at the mangrove area. 
A second site was 1 5 km southeast at the coastal 
fishing wharf of the town of Parit Jawa. This site 
was used for field observations of foraging behav- 
ior and as a supply of animals for the laboratory 
observations. Although three species of homal- 
opsine snakes were encountered in the study areas 
(Cerberus rynchops, Bitia hydroides, and Fordonia 
leucobalia), C. rynchops was the most common 
species and is the subject of this report. 

Initially, sampling along the shoreline at the pri- 
mary site was conducted at all tidal stages and at 
sunrise, mid-morning, noon, mid-afternoon, sun- 
set, and at night. Individuals of C. rynchops were 
observed during all tidal stages and times of day, 
but the snakes were most active on the surface of 
the mud flat, on the beach, and entering and leav- 
ing the stone wall from 1900-2200. During this 
time, snakes also were commonly on the leading 
edge of the incoming tide, and most of our col- 
lecting concentrated on this period. Notes on the 
time of capture, habitat, tidal stage, and water depth 
were recorded at the time of capture. 

Following two to three hours of collecting, snakes 
were taken to our laboratory and palpated for 
stomach contents. If something in the stomach was 
detected by palpation but was not regurgitated, 



then the snake was preserved and the contents 
were removed by dissection. The remaining snakes 
which failed to regurgitate any food were assumed 
to have empty stomachs (this assumption is sup- 
ported by the dissection of a series of 24 snakes 
early in this work). Snout- vent length (sv) and tail 
length were then measured to the nearest 0.5 cm, 
and the snake was tagged (Floy Tag & Mfg. Inc., 
#FD-67C). Mass of the snake (Ms) was determined 
to the nearest gram by confining the snake with a 
plastic bag and placing it on a digital top-loading 
scale. The snakes were released within an hour of 
measurement at the shoreline in the center of the 
study area. 

Prey items were immediately preserved using 
10% formalin. After fixation, the maximum di- 
ameter and various lengths of the fish were mea- 
sured to the nearest 0. 1 mm, and after excess fluid 
was blotted off, the mass of the fish (Mf) was de- 
termined to the nearest 0. 1 g. Based on preserved 
series of fish, linear least squares regressions es- 
timated log Mf from the log of various measures 
of length from partially digested fish. 

Three methods were used to estimate the pro- 
portion of diet comprised by each fish species. The 
occurrence of fish species was estimated as the 
percentage of stomachs in which they occurred as 
well as the percentage of the total number of items 
they comprised. The original wet biomass of each 
species was also estimated and then expressed as 
a percentage of the estimated total biomass of all 
prey consumed by one sample. 

Observations were made on the feeding of fresh- 
ly captured C. rynchops from 3 1 January-5 March 
1984 and 10 December 1985-31 January 1986. 
Captive snakes were kept and observed feeding in 
1.5 cm of fresh water inside white Styrofoam con- 
tainers with an inside height, length, and width of 
35, 40, and 60 cm, respectively. The snakes were 
observed feeding at times ranging from 1 1 00-2300, 
with air temperatures ranging from 26°-32°C. 
Snakes that failed to feed at three consecutive trials 
were released. A total of 23 snakes (sv 29.5-60.0 
cm) was observed feeding for 34 trials, during which 
7 1 mudskippers (Periophthalmus chrysospilos) were 
consumed. The times between trials ranged from 
two to four days. For 1 9 trials involving 1 2 snakes, 
a single fish was offered to the snake. For the re- 
maining trials, two to five fish were offered in rapid 
succession. No more than five feedings were ob- 
served per snake. Videotape facilitated the docu- 
mentation of the 40 feedings during 1985-1986. 

The total length (tl) of each fish was measured 
to the nearest millimeter immediately before of- 



FIELDIANA: ZOOLOGY 




LEGEND: 



WATER 






if']!?'- MUD FLATS 






MANGROVES 



SUMATRA 
(INDONESIA) 



SINGAPORE/ 



104* 



Fig. 1 . Schematic map of the study site in the Muar estuary. Areas indicating mud flat estimate the extent of 
exposed mud during the spring low tides. 



JAYNE ET AL.: CERBERUS RYNCHOPS 



fering it to the snake. The tl ranged from 32-88 
mm for the fish used in the experiments. The Mf 
were estimated to the nearest 0. 1 g by a linear least 
squares regression (log[MfJ - 3.543[log(TL)] — 
4.999, r 2 = .983) which was calculated from the 
measurements of 24 fish that had been preserved 
in 10% formalin. For each trial, the initial position 
(ip) where the snake seized the fish was assigned a 
value from 1 to 4 as follows: 1 = the mouth of the 
fish was somewhere within the mouth of the snake, 
2 = the snake bit the head or gill region while the 
mouth of the fish remained outside of the snake's 
mouth, 3 = the fish was seized between the pos- 
terior margin of the operculum and the posterior 
end of the first dorsal fin, and 4 = the fish was 
seized posterior to the first dorsal fin. The amount 
of struggle (s) by the fish after being seized was 
rated subjectively from 1 to 3 as follows: 1 = the 
fish only displayed slight movement when it was 
initially seized, 2 = the fish struggled slightly while 
it was being held, and 3 = the fish struggled more 
than twice and violently enough to substantially 
move the head and neck of the snake. 

A digital stopwatch was used to determine (to 
the nearest second) the total prey handling time 
(Tt). Three phases were timed: holding time (Th), 
the initial seizure of the fish until the start of lateral 
jaw walking by the snake; jaw-walking time (Tjw), 
the start of lateral jaw walking until the snout of 
the fish was in the mouth of the snake; and swal- 
lowing time (Ts), the entrance of the snout of the 
fish into the mouth of the snake until the end of 
swallowing as indicated by the disappearance of 
the tail of the fish. 

These various times of prey handling were ana- 
lyzed as the dependent variable in three multiple 
regression models that were calculated using a hy- 
brid stepwise procedure with anF> 4.0 (P < .05) 
as the criterion for addition or deletion from the 
model. Model 1 estimated Tt by summing sepa- 
rate regression estimates of Th, Tjw, and Ts. Oc- 
casionally, if a negative value was predicted for 
one of these three times, this estimate was changed 
to zero before adding it to the other times con- 
tributing to Tt. Models 2 and 3 each used a single 
multiple regression equation to predict Tt. Models 
1 and 2 used a partial F > 4.0 (P < .05), whereas 
model 3 used partial F = 2.9 (.05 < P < .10) to 
determine the independent variables to be includ- 
ed in the multiple regression model. For all han- 
dling times sv, Ms, tl, Mf, ip, s, and ordinal num- 
ber of fish within a trial were used as the 
independent variables. For Tjw and Ts, Th was 
also used as an independent variable. 

Some Cerberus rynchops from Muar and Parit 



Jawa were brought back to the United States for 
additional experiments. These snakes were main- 
tained on a diet of live goldfish and were usually 
fed weekly in 1.5 cm of fresh water inside of a 
Plexiglas aquarium (75 cm long x 50 cm wide x 
30 cm high). A 2-cm grid on white paper under- 
neath the aquarium provided fixed points of ref- 
erence. 

Two experiments were conducted to clarify what 
stimulus might facilitate prey capture for C. ryn- 
chops. For both experiments the water tempera- 
ture was from 28°-29°C, and the snakes were given 
five minutes to acclimate to the aquarium before 
being tested. In the first experiment, nine goldfish 
were placed in the aquarium for 30 minutes and 
then removed immediately prior to the introduc- 
tion of a snake. At intervals of about 30 seconds, 
0.5-ml samples of aquarium water were dropped 
into the aquarium from a hypodermic syringe held 
30 cm above the water. The grid was used to es- 
timate the distance between the head of the snake 
and landing point of the drops. The water was 
dropped only when all of the lower jaw of the snake 
was below the water's surface, allowing the drops 
to land laterally and anteriorly to the head of the 
snake. Four snakes were subjected to this stimulus 
ten times in succession, with two of the snakes 
receiving drops within 4-8 cm the first five times 
and 8-12 cm the second five times, and the other 
two snakes receiving drops in the reverse order. 
In the 2nd experiment, the same four snakes were 
subjected to a similar procedure with clean water 
in the aquarium. The aquarium was rinsed out ten 
times between trials. Results were only kept for 
snakes which ate a goldfish within five minutes of 
the conclusion of an experiment. Videotape re- 
corded the orientation of the strike during some 
other regular feeding sessions with goldfish as prey. 



Results 
Diet 

Of the 262 Cerberus rynchops collected and pal- 
pated in 1984, 97 had one or more prey items. In 
nine cases, however, the stomach contents were 
fed back to the snakes that had been captured for 
a second time; consequently, only the prey items 
of the remaining 88 snakes were analyzed. Of the 
349 C. rynchops collected in 1985-1986, stomach 
contents were obtained from 93, and all of these 
items were retained for analysis. 

A total of 3 1 3 items was removed from the 1 8 1 



FIELDIANA: ZOOLOGY 



Table 1 . Size distribution for samples of Cerberus rynchops. 











Snake size 


classes by snout-vent length (cm) 








25- 


30- 


35- 


40- 


45- 


50- 


55- 


60- 


65- 


Sample 


n 


29.5 


34.5 


39.5 


44.5 


49.5 


54.5 


59.5 


64.5 


69.5 


1984 diet 


88 


4.5 


27.3 


48.9 


14.8 


4.5 










1985-1986 diet 


93 


1.1 


15.0 


22.5 


18.3 


17.2 


11.8 


5.4 


5.4 


3.2 


15 Jan.-16 Feb. 1984 


181 


7.7 


28.2 


43.6 


11.6 


5.5 


2.2 


0.5 


0.5 




1-5 March 1984 


67 


3.0 


20.9 


41.8 


23.9 


5.9 


4.5 








20 Nov.- 17 Dec. 1985 


237 


5.5 


21.9 


29.5 


13.9 


10.5 


3.4 


2.5 


2.1 


0.4 


18 Jan.-8 Feb. 1986 


112 


5.4 


17.0 


19.6 


16.1 


12.5 


17.0 


4.5 


7.1 


0.9 



Frequencies of occurrence are all given in percentages for each 5-cm size class within a sample. 
The first two rows in the table indicate the snakes with prey items used for analysis of diet. The remaining rows 
indicate the size distribution of all snakes collected between the dates indicated at left, 
n = Sample size. 



C. rynchops with stomach contents. These snakes 
ranged from 26-67 cm in sv and from 19-208 g 
in Ms. The size distribution of snakes with stom- 
ach contents is summarized in Table 1. The stom- 
achs of all these snakes mostly contained four 
species of oxydercine gobies, including 2 1 5 Oxu- 
derces dentatus, 27 Scartelaos pectinirostris, 13 
Periophthalmus chrysospilos, and 2 Boleophthal- 
mus boddarti. Other gobiidae found in C. rynchops 
included six large, elongate fish (5 Taeniodes cir- 
ratus and 1 Odentamblyopus rubicundus) and two 
small fish of the genus Acentrogobius. Two species 
of catfish were represented by 1 9 specimens of 
Ariidae (Arius sp.) and one Plotosidae (Plotosus 
sp.). Twelve mullet (Mugilidae) were consumed, 
of which four were identifiable as Liza sp. and two 
as Valamugil sp. Two tongue fish (Cynoglossidae, 
Cynoglossus sp.) and two Sillaginidae (Sillago sp.) 
were also eaten. Only a single eel (Synbranchidae, 
Macrotema sp.) was removed from C. rynchops. 
One specimen each of Eleotrididae (Butis sp.) and 
Polynemidae (Eleutheronema tetradactylum) was 
also found. The remaining nine fish recovered from 
C. rynchops were not identifiable. 

Of the 181 C. rynchops, slightly more than half 
(109) had only one item in the stomach. Of the 
remaining 72 snakes with multiple prey items, 38 
had 2 items, 2 1 had 3 items, 5 had 4 items, 5 had 
5 items, and 1 snake each had 6, 7, and 8 items; 
25 of these snakes had taken only Oxuderces den- 
tatus. 

The size distribution of snakes with stomach 
contents collected in 1984 was significantly dif- 
ferent from that of the 1985-1986 sample (x 2 = 
43.44, df=S,P< .001). Compared to snakes 
collected in 1985-1986, those sampled in 1984 
had proportionately fewer individuals with sv > 
45 cm (4.5% vs. 43%). To facilitate comparisons 
between snakes of the two study periods, an sv of 



45 cm was used to subdivide samples. Table 2 
summarizes the percentage of diet comprised by 
the major groups of prey species for small and large 
snakes. The 1984 and 1985-1 986 samples of small 
snakes (sv < 45 cm) are very similar, with Oxu- 
derces dentatus comprising the largest portion of 
diet using any of the three measures of importance. 
The diet of the large C rynchops (sv > 45 cm) 
differs markedly from that of the smaller snakes. 
Although O. dentatus comprised the greatest por- 
tion of items in the larger snakes, it accounted for 
less than 10% of the biomass consumed. Equal 
percentages of large snakes contained O. dentatus 
and sea catfish, but the sea catfish had almost twice 
the biomass of the O. dentatus. Together, mullet 
and elongate gobies comprised less than 20% of 
the prey items; however, they accounted for nearly 
two-thirds of the prey biomass of the large snakes. 

Predator/Prey Size Relationships 

Total mass of the stomach contents per snake 
significantly increased with the Ms (fig. 2). These 
data were log transformed to equalize variance of 
the dependent variable. For the 181 snakes, log 
total mass of contents consumed per snake = 
-1.198 + 0.875 log Ms; r 2 = .24. For example, 
this least squares regression predicts a 50-g snake 
would consume 1 .94 g, about 4% of the Ms. These 
predicted masses of meals are much less than the 
maximum consumed by snakes. For example, one 
Cerberus rynchops (sv = 64 cm, Ms = 124 g) con- 
sumed a single mullet (Liza sp.; maximum height 
x width = 37 x 26 mm; Mf = 66 g) that was 53% 
of the Ms. However, meals of such large relative 
size were uncommon for the snakes sampled in 
this study. In fact, the second largest meal was 
another Liza sp., and it comprised only 28.8% of 
the Ms. Only 25 of the 181 snakes with contents 



JAYNE ET AL.: CERBERUS RYNCHOPS 



Table 2. Diet of Cerberus rynchops. 





n 






Percentage 


occurrence 


of prey species 






Sample 


Od 


Sp 


Pc 


Ar 


M 


EG 


Other 


1984 <45 cm 


















% Snakes 


84 


77.4 


16.7 


6.0 


1.2 


3.6 


1.2 


9.5 


% Total items 


155 


79.4 


9.0 


3.2 


0.6 


1.2 


0.6 


1.2 


% Prey biomass 


















(total 131.9 g) 




64.6 


13.0 


2.0 


1.7 


7.0 


6.4 


5.6 


1985-1986 <45cm 


















% Snakes 


53 


67.9 


17.0 


9.4 


9.4 


3.8 





15.1 


% Total items 


93 


65.6 


9.7 


5.4 


5.4 


2.2 





11.8 


% Prey biomass 


















(total 103.9 g) 




52.9 


8.1 


7.4 


8.9 


1.4 





21.2 


1985-1986 >45cm 


















% Snakes 


40 


32.5 


7.5 


5.0 


32.5 


17.5 


12.5 


2.5 


% Total items 


58 


43.1 


5.2 


5.2 


22.4 


12.1 


8.6 


3.4 


% Prey biomass 


















(total 271.5 g) 




8.8 


1.2 


1.5 


15.6 


48.8 


22.6 


1.5 



Od = Oxuderces dentatus; Sp = Scortelaos pectinirostris; Pc = Periophthalmus chrysospilos; Ar = ariid catfish; 
M = mullet; and EG = elongate gobies. See text for complete explanation of prey categories. 

Percentage of snakes with prey species does not sum to 100 for a sample because of stomachs containing more 
than one species. 



had relative mass of the total contents > 10%. The 
snakes with the six largest relative masses of stom- 
ach contents each had consumed single fish, none 
of which were oxydercine gobies. The seventh larg- 
est set of contents consisted of three Oxuderces 
dentatus which were 18.4% of the mass of a 25- 
cm snake. Snake sv did not significantly affect the 
number offish consumed (F — .40, df= 1,179; P 
> .50). 

A detailed comparison of the size of the prey 
relative to the morphological limits of gape is be- 
yond the scope of this study, but some evidence 
suggests that C. rynchops tends to take relatively 
small prey. Although the shape of fish may vary 
radically among different taxa, the maximum di- 
ameter of a fish approximates the difficulty a snake 
may have swallowing it. In addition to the mullet 
mentioned previously, some of the largest maxi- 
mum diameters of fish consumed by C. rynchops 
were 13.0, 19.7, 19.9, and 31.7 mm for snakes 
with sv of 27, 38, 46, and 62 cm, respectively. In 
contrast to these large fish, 7.9 mm was the largest 
maximum diameter measured for any of the 2 1 6 
O. dentatus consumed by C. rynchops. 



Foraging 

Water conditions at the Parit Jawa site often 
permitted observation of Cerberus rynchops for- 
aging in water as deep as 1.3 m. Whether in water 



or on mud fiat, snakes were rarely sedentary for 
more than a minute. Swimming C. rynchops con- 
sistently moved along the bottom in contrast to 
the surface swimming that is commonly used by 
colubrid snakes such as Nerodia (Jayne, 1985). 
Snakes usually performed sidewinding locomo- 
tion on mud that was firm enough to support their 
weight. If snakes sank in mud past the first few 
dorsal scale rows, then lateral undulation was used 
for surface locomotion as well as swimming through 
the mud slightly below its surface. Snakes usually 
explored burrows and irregularities of the sub- 
strate regardless of whether they were under water. 
Occasionally, snakes swam with their mouths open 
slightly, and the lateral movements of the head 
were exaggerated compared to that during normal 
swimming. On two of these occasions, individuals 
of C. rynchops were observed capturing very small, 
schooling fish, and two other snakes used this be- 
havior to capture an Oxuderces dentatus and a 
mullet that had just escaped after the snake at- 
tempted to swallow it. In two other instances, 
snakes swimming in muddy water were observed 
with this open-mouthed posture, but no fish could 
be seen. Another snake remained stationary, as it 
was in the midst of a school of fish, and it re- 
peatedly used similar alternating lateral move- 
ments of the head and neck until the school offish 
dissipated. Two other strikes at fish observed in 
the field also seemed to have a distinct lateral com- 
ponent. 



FIELDIANA: ZOOLOGY 



2.0 



1.5 



E 

3 1.0 

-i 
< 

2 



rrt 0.5 

to 

<0 

< 

< 
i- 
O 

I- 

o 
o 



-0.5 



•1.0 



-1.5 




1.0 



1.5 2.0 

LOQ SNAKE MASS (gm) 



2.5 



Fig. 2. Plot of log total mass of stomach contents versus log mass of snake for the 181 Cerberus rynchops with 
prey items. Both masses were originally in grams. The line indicates the least squares regression, where log mass of 
stomach contents = - 1.198 + 0.875 log snake mass, r 2 = .24. 



Analysis of video tapes of 65 strikes of captive- 
fed C. rynchops confirmed that there was always 
a lateral movement involved in aquatic prey sei- 
zure (fig. 3). The initial phase of the strike could 
be directed in nearly any direction; however, a 
subsequent rapid lateral flexion of the neck mo- 
mentarily caused a posture with the anterior region 
of the snake forming an arc of about 270° (fig. 3). 
This quick lateral flexion usually occurred just as 
the snake's mouth contacted the fish. During this 
stage of prey seizure, the fish would often not be 
grasped securely in the snake's jaws, and the ori- 
entation of the snake frequently trapped the fish 
between the snake's mouth and body. This en- 
abled some snakes to quickly reposition their jaws 
or to recapture fish that had momentarily escaped. 



Feeding Behavior 

The following descriptions are representative of 
the variation in observed captive feeding behav- 
ior. The figures in parentheses indicate the elapsed 
time (in seconds) after the snake initially seized 
the fish. 

A Cerberus rynchops (sv = 32 cm, Ms = 22 g) 
seized a Periophthalmus chrysospilos (tl =75 mm, 
Mf = 4.7 g) just posterior to the operculum as the 
fish was moving near the snake. Immediately after 
striking the fish, the snake rapidly moved the fish 
back to the corners of its mouth and held the fish 
perpendicular to its neck. During this initial sei- 
zure, the fish moved only slightly. As the snake 
continued to hold the fish, there were occasional 



JAYNE ET AL.: CERBERUS RYNCHOPS 




Fig. 3. Striking behavior of Cerberus rynchops attacking a goldfish. The illustration is based on tracings made 
from videotape. Pairs of successive images are superimposed, with the dotted outline indicating the earlier position 
in each pair. A, Position at time = and 1/15 second; B, position at time =1/15 and 2/15 second. 



FIELDIANA: ZOOLOGY 



Table 3. Coefficients for multiple regression equations predicting prey handling times for Cerberus rynchops 
consuming Periophthalmus chrysospilos. 



Dependent 




Coefficients of independent variables 




Constant 

(sec) 




variable 

(sec) 


Mf 

(sec/g) 


sv 
(sec/cm) 


s 
(sec) 


ip 
(sec) 


Multiple 
r 2 


Model 1 
Th 
Tjw 
Ts 


62.1 (.37) 
12.4 (.52) 
11.3 (.45) 


-7.79 (-.28) 
-1.13 (-.29) 
-1.78 (-.43) 


105 (.31) 

NS 
NS 


NS 

9.29 (.31) 

NS 


117 
18 
79 


.43 
.50 
.35 


Model 2 
Tt 


84.1 (.43) 


-10.1 (-.32) 


125 (.32) 


NS 


185 


.53 


Model 3 
Tt 


83.0 (.43) 


-10.7 (-.34) 


94 (.24) 


39* (.16) 


166 


.55 



Figures in parentheses after coefficients are standardized regression coefficients. 

n = 7 1 for all regressions; ns = not significant; sv = snout-vent length of snake; Mf = mass of fish; ip = initial 
position where snake seized fish; and s = struggle of fish. 
♦Partial F = 2.95 (.05 < P < .10). 



biting-like movements of the snake's maxillae (20, 
278, and 392). While the fish was being held by 
the snake, some fin, gill, and mouth movements 
were apparent. The snake then began to lateral jaw 
walk toward the snout of the fish (400) while the 
fish showed only very slight gill and mouth move- 
ments. Immediately after reaching the snout of the 
fish (507), swallowing began and continued until 
the tail of the fish disappeared from view (590). 
As the snake was swallowing, no fish movements 
could be discerned. 

The duration of this holding behavior by C. 
rynchops varied considerably as illustrated by 
another individual (sv = 36.5 cm, Ms = 35 g) that 
ate a mudskipper (tl = 69 mm, Mf = 3.5 g). This 
snake seized the fish on the gill region, and the fish 
flopped violently as the snake briefly held it. The 
snake started slow lateral jaw walking to the snout 
of the fish (9) as the fish continued to make whole 
body undulations. Upon reaching the snout of the 
fish (63), the snake started swallowing, and the fish 
continued to move slightly until its tail disap- 
peared from view (94). 

During another trial, a snake (sv = 34 cm) was 
disturbed and released the fish (2.5-g mudskipper) 
after holding it for 117 seconds. The snake was 
then removed from the container, and the fish was 
observed until it died 16.5 minutes after being 
seized. During other feeding trials with mudskip- 
pers, as the snake held the fish, there was some- 
times a marked darkening of the fish that spread 
from the site of the bite. Occasionally, there was 
also a noticeable dilation of the pupils of the mud- 
skippers while they were being held. In the field, 
a C. rynchops was observed holding a mullet (tl 



= 82 mm, Mf = 7.2 g) that was still moving slight- 
ly. By the time the snake was captured, the fish 
had been released and had died. In the laboratory, 
several snakes (sv = 29-51 cm) were forced to 
release goldfish (0.8-3.7 g) just as lateral jaw walk- 
ing began. The Th varied from 0.2-6.9 minutes. 
Of the 3 1 observed goldfish, 1 6 died after being 
held from 1.1-6.9 minutes. The times of death 
after initial seizure ranged from 6.0-44.0 minutes; 
nine of these 16 goldfish died in less than 16.5 
minutes after being seized by snakes. Hence, the 
venom of C. rynchops appears capable of immo- 
bilizing and killing selected prey. 

Table 3 summarizes the coefficients of the sig- 
nificant independent variables in the various mul- 
tiple regression equations. As suggested by the 
standardized regression coefficients, the Mf was 
always the most significant factor affecting all prey 
handling times. Increased sv of the snake always 
significantly decreased handling times. Struggling 
by the fish primarily increased Th. More posterior 
ip increased predicted Tjw. Interestingly, Th (and 
presumed envenomation) did not significantly af- 
fect Tjw or Ts. 

Table 4 lists select predicted values for the three 
models of total handling time. The Ms can be 
predicted from sv by the least squares regression 
log Ms = 2.878(log sv) - 3.0 1 8, r = .969, n = 1 8 1 . 
A 45-cm C. rynchops has about twice the mass of 
a 35-cm snake (55 vs. 27 g). For a given size, s, 
and ip of mudskipper, predicted Tt for the 45-cm 
snake can be from '/j-% that predicted for the 35- 
cm snake. For a given snake, handling a 2-g mud- 
skipper may take from V-tr- x k the Tt predicted for 
a 4-g fish. Increased struggle of the fish may cause 



JAYNE ET AL.: CERBERUS RYNCHOPS 



Table 4. Predicted handling times (in seconds) for Cerberus rynchops eating Periophthalmus chrysospilos (see text 
for explanation of models). 





Independent variable 






Model 1 




Model 2 




Mf 


sv 






Model 3 


(g) 


(cm) 


s 


ip 


Th 


Tjw 


Ts 


Tt 


Tt 


Tt 


2 


35 


1 


3 


74 


31 


39 


144 


124 


169 


4 


35 


1 


3 


198 


54 


62 


314 


208 


252 


2 


35 


1 


2 


74 


21 


39 


134 


208 


213 


4 


35 


1 


2 


198 


47 


62 


307 


292 


379 


4 


45 


1 


3 


119 


44 


44 


198 


191 


247 


4 


45 


1 


2 


119 


35 


44 


189 


191 


208 


2 


45 


1 


3 


0* 


19 


21 


40 


23 


42 


2 


35 


3 


3 


284 


30 


39 


353 


374 


238 


2 


45 


3 


3 


206 


19 


21 


246 


273 


240 


2 


45 


1 


1 


0* 


1 


21 


22 


23 


-182 


4 


55 


1 


3 


42 


33 


26 


101 


91 


120 


4 


55 


3 


3 


252 


33 


26 


311 


340 


318 


2 


55 


1 


3 


0* 


9 


4 


13 


-78 


-36 



* Negative value was changed to 0. 

Mf = Mass offish; sv = snout-vent length of snake; s = struggle offish; and ip = initial position where snake seized 
fish. 



up to a sixfold increase in Tt and elicit holding 
behavior as well. 

The data from the stimulus experiments were 
tallied as strike or no response, combined for all 
four of the snakes (n = 40), and arranged into two- 
by-two contingency tables for chi-squared analysis 
(x 2 = 3.84, P < .05 used for decision-making). For 
the experiment using water with fish odor, 10 of 
the 16 strikes occurred during the first half of each 
trial using clean water. Hence, for water with fish 
odor (x 2 = 1.67) and for clean water (x 2 = 0), 
response does not appear to be dependent on the 
number of stimuli within each trial. In other words, 
the snakes did not appear to be habituating to the 
ten stimuli within each trial. When using the water 
with fish odor, 15 strikes resulted from stimulus 
within 4-8 cm of the head of the snake, and only 
one strike occurred for the 8-12-cm distance; 
therefore, response was dependent on the distance 
from the stimulus (x 2 = 20.67). For the experiment 
with clean water, 1 2 strikes were within 4-8 cm, 
and six strikes were within 8-12 cm. The x 2 was 
equal to 3.64, just slightly less than the critical 
value. During all of the experiments and routine 
feeding sessions, on only two occasions did snakes 
attempt to strike at a handler or at moving objects 
above the surface of the water. Thus, the response 
to the waterdrop stimulus does not appear to be 
defensive or visual in nature. Instead, this re- 
sponse appears to be predatory and largely the 
result of tactile stimulus. 



Growth 

As indicated by a high incidence of zero and 
negative growth of 35 snakes recaptured in 1984 
less than 20 days after marking, short-term growth 
was probably obscured by measurement error and 
handling stress. Consequently, the samples ana- 
lyzed here are confined to 24 snakes recaptured 
after 20 or more days. Table 5 lists relative growth 



Table 5. Snout- vent lengths and masses for 14 of 
the 24 recaptured Cerberus rynchops with the greatest 
percentage increase in mass. 



Snake 


Elapsed 


Initial 


Initial 


no. 


days 


sv (cm) 


Ms(g) 


628 


20 


32.5(1.5%) 


19(15.8%) 


630 


20 


38.0 (3.9%) 


35 (28.6%) 


3103 


25 


43.0(1.2%) 


50(14.0%) 


3090 


26 


30.5(1.6%) 


16(18.8%) 


3073 


27 


38.0(1.3%) 


29 (20.7%) 


3074 


28 


37.0 (2.7%) 


28 (14.3%) 


3094 


28 


42.0(1.2%) 


42(21.4%) 


3045 


28 


49.0 (2.0%) 


70(17.1%) 


630 


29 


38.0 (3.9%) 


35(17.1%) 


3023 


32 


40.0 (8.8%) 


41 (17.1%) 


1091* 


42 


40.5 (6.0%) 


39 (25.6%) 


1030* 


45 


50.0 (2.0%) 


67 (23.9%) 


1038* 


65 


44.5 (3.8%) 


51 (13.7%) 


922* 


68 


60.0 (5.0%) 


124(8.1%) 



Figures in parentheses indicate percentage increase be- 
tween initial and final capture. 
* Captured during 1986. 



10 



FIELDIANA: ZOOLOGY 



for some of these snakes with the greatest increase 
in Ms. At initial capture, the sv of the 1984 sample 
of 15 snakes ranged from 30.5-49.0 cm (x = 38.5 
cm, 5 = 5.04), and the Ms, from 16-70 g (x = 
36.67 g, 5= 15.31). The average elapsed time be- 
tween captures for this group was 26.7 days. 
Growth varied considerably; on average, these 
snakes gained mass at 0.50%/day (range -0. 19%- 
0.76%/day, 5 = 0.40) and grew in sv at 0. 1 1%/day 
(range 0%-0. 1 9%/day, s = 0.075). The nine snakes 
recaptured in 1 986 initially ranged from 40.5-60.0 
cm sv (x = 49.3 cm, s = 19.3) and from 32-124 
g (x = 54.4 g, 5 = 28.1). Average time between 
captures was 44.8 days (s = 6.5) for this group. 
Average growth rates for the 1986 recaptures were 
0. 1 2%/day (range -0.28%-0.6 1 %/day, s = 0. 1 5 1) 
for Ms and 0.06%/day (range 0%-0.27%/day, 5 = 
0.04) for sv. For both samples combined, average 
growth rates were 0.36%/day (s = 0.40) for Ms and 
0.09%/day (s = 0.07) for sv. 



Population Structure 

Table 4 lists the distributions of snake sv for 
two subsamples each for 1984 and 1985-1986. 
Using a chi-square test, no significant differences 
were found between the two subsamples within 
1984 (x 2 = 1 1.47, df= 7, .1 < P < .2). Similarly, 
no differences in size distribution were evident 
when comparing the 1985 to the 1986 subsamples 
(X 2 - 13.30, df= 9, .1 < P < .2). This and the 
fact that small snakes (sv < 30 cm) were contin- 
uously encountered during this study suggest that 
reproduction of this population is aseasonal. When 
the total size distribution of 1984 was compared 
to that of 1985-1986, a highly significant differ- 
ence was found (x 2 = 77.18, df=9,P<^ .001). 



1978; Voris & Jayne, 1979; Voris, 1985), but rare- 
ly trapped C. rynchops. These observations and 
the high concentrations of subadults encountered 
in the study area lead us to believe that we could 
estimate the subadult population in the study area 
within a limited period of time. 

Three estimates were made. For the first esti- 
mate, 108 snakes were marked and released be- 
tween 15 January and 10 February 1984. Collect- 
ing on 12-13 February produced 32 unmarked 
snakes and 12 previously marked snakes. Using 
Bailey's (1952) formula the population size esti- 
mate is 374 (s = 84.3). For the second estimate, 
the snakes collected on 12-13 February and two 
other snakes collected earlier were marked and 
released. The population was not disturbed by us 
from 15 February-1 March. From 1-5 March, we 
collected 44 unmarked snakes and 2 1 marked pre- 
vious to 1 5 February. Bailey's estimate for these 
data is 426 (s = 72.5). In 1985-1986 the third 
estimate was made. From 20 November-17 De- 
cember 1985, 210 snakes were marked. The snakes 
were left undisturbed until 1 8 January-8 February 
1986, whereupon 1 12 animals were collected. Of 
the 16 recaptures during this period, seven had 
Floy tags, and the rest had conspicuous scars where 
the tags had pulled out. Bailey's estimate for this 
period was 1,396 (s = 303). 

During the strongest tides, the area of the in- 
tertidal zone within the study site is about 80,000 
m 2 . The conspicuous concentration of snakes at 
the edge of the water and the unknown extent to 
which deeper water is utilized by the snakes, how- 
ever, complicate calculation of the density per unit 
area attained by C. rynchops at this site. Never- 
theless, these estimates of population size suggest 
there may be from one to three subadult snakes 
per meter of shoreline within the primary study 
site. 



Population Size Estimates 



Although most collecting was confined to hab- 
itats within the primary study site, two adjacent 
habitats were investigated. On the landward side 
of the beach and stone wall, there was a mowed 
soccer field and an unmowed grass field with a 
large freshwater pond. No Cerberus rynchops were 
observed in about six man-hours of exploring and 
traversing this area. The portion of the river mouth 
below the low tide level and about 100 r.i north 
of the east end of the study site is serviced by two 
stake nets. These nets have produced extensive 
collections of sea snakes since 1975 (Voris et al., 



Discussion 

For the communities of marine snakes that have 
been previously studied, little or no overlap in diet 
has been found. For a community of ten hydro- 
phiids on the Ashmore reef in Australia, Mc- 
Cosker ( 1 975) found practically no overlap in either 
the diet or microhabitat preferences of the different 
species. Similarly, for four different communities 
of acrochordids and hydrophiids in Malaysia, Voris 
and Voris (1983) found most species were dietary 
specialists, and only modest overlap occurred 



JAYNE ET AL.: CERBERUS RYNCHOPS 



11 



among the more dominant species of the com- 
munity. Lapemis hardwickii is a notable exception 
to this trend, as this hydrophiid has a very gen- 
eralized diet (Voris & Voris, 1983). 

In the Muar estuary, the homalopsine Fordonia 
leucobalia feeds exclusively on crabs and has no 
dietary overlap with other snakes. Preliminary 
analysis of the diet of Bit ia hydroides suggests this 
homalopsine feeds primarily on gobies and hence 
has overlap with the diet of Cerberus rynchops. 
Acrochordus granulatus captured from the Straits 
of Malacca consume about 46% Eleotrididae and 
54% Gobiodei with taenioid gobies comprising 
7.7% of the prey items (Glodek & Voris, 1982). 

The diets of juvenile and adult Enhydrina schis- 
tosa are comparable, and this species, which is the 
most abundant hydrophiid at Muar, consumes 
76.7% ariid and 13.8% plotosid catfish (Voris et 
al., 1978). The second most abundant hydrophiid 
at Muar {Hydrophis melanosoma) eats exclusively 
eels (Glodek & Voris, 1982). Various gobies com- 
prise about 10% of the prey items of the third most 
abundant hydrophiid {Hydrophis brookii) at Muar. 
Hydrophis torquatus is the only other hydrophiid 
at Muar for which dietary information is available, 
and small samples suggest this species consumes 
60% taenioid gobies (Glodek & Voris, 1982). 

The extent of diet overlap can be calculated us- 
ing the Schoener (1968) index, ex. For the 1985- 
1986 sample of large C. rynchops compared with 
Enhydrina schistosa, oc = .17, whereas overlap 
between E. schistosa and 1985-1986 small C. ryn- 
chops was only .05. Using the species level for 
grouping prey items, no overlap occurred between 
C rynchops and either H. melanosoma or H. 
brooki. For large C. rynchops compared with H. 
torquatus and A. granulatus, oc = .07. For the more 
abundant snake species within a community, Glo- 
dek and Voris (1 982) found oc rarely exceeded .10. 

The extent to which dietary overlap is deter- 
mined by predator choice versus microhabitat 
preferences remains unclear. During all of the col- 
lecting of homalopsines at Muar and Parit Jawa, 
not a single hydrophiid was seen. The extensive 
use of fishing nets has captured hundreds of hy- 
drophiids in the main channel of the Muar River 
(Voris et al., 1 978); however, these same nets have 
yielded less than ten homalopsines. The relative 
scarcity of adult C. rynchops collected from the 
tidal edge and the occurrence of prey such as Tae- 
niodes cirratus imply that large individuals of C. 
rynchops are more likely to occur in deeper water 
than small individuals. Unfortunately, it is diffi- 
cult to collect snakes in this most probable region 



of interspecific spatial overlap at water depths 
ranging from 1-3 m. Yet it seems likely that the 
greater dietary overlap of E. schistosa and large C. 
rynchops is primarily the result of ontogenetic 
changes in habitat preference which cause the rel- 
atively opportunistic C. rynchops to overlap more 
with the more specialized diet of E. schistosa. 



Feeding Behavior 

Aspects of the feeding behavior of Cerberus ryn- 
chops, such as prey detection, capture, and han- 
dling, resemble those of other aquatic snakes. Cer- 
berus rynchops used a predominately lateral strike 
to capture prey. Pelamis platurus is a surface feed- 
ing hydrophiid, and it also uses a lateral strike to 
capture fish (Pickwell, 1972; Kropach, 1975). 
Another hydrophiid, Enhydrina schistosa, feeds 
primarily along the bottom and it also uses a lateral 
strike to capture fish. Both P. platurus and E. schis- 
tosa hold and envenomate fish and wait until 
struggling ceases before swallowing (Pickwell, 1 972; 
Voris et al., 1978). As shown in this study, C. 
rynchops were more likely to hold (and presum- 
ably envenomate) fish that were relatively large or 
struggled vigorously. However, initiation of swal- 
lowing by C. rynchops may or may not occur be- 
fore the fish has stopped struggling. Despite the 
sharp spines present in the dorsal and pectoral fins 
of ariid catfish, some individuals of C. rynchops 
in the field were observed swallowing these catfish 
while they were still moving. Catfish are always 
consumed head first by E. schistosa (Voris et al., 
1978) and by individuals of C. rynchops observed 
in this study. As one might expect for snakes that 
inhabit muddy water and have nocturnal tenden- 
cies, C. rynchops readily showed striking behavior 
when exposed to mechanical stimulus. Feeding of 
P. platurus also appears responsive to mechanical 
stimulus (Kropach, 1975). As evidenced by the 
ability of E. schistosa to feed in total darkness, 
some combination of tactile and olfactory cues 
appear sufficient for prey capture and consump- 
tion (Voris et al., 1978). 

In a series of carefully controlled experiments, 
Drummond (1979, 1 985) has examined the effects 
of visual and olfactory stimuli on predatory be- 
havior of certain piscivorous natricine snakes. 
Drummond (1979) found that individuals of Ner- 
odia sipedon were not entirely dependent on chem- 
ical cues to locate and capture fish. Moving models 
offish were more effective than nonmoving models 
for eliciting orientation, attacking, and searching 



12 



FIELDIANA: ZOOLOGY 



behavior by N. sipedon. Among the predatory be- 
haviors described for N. sipedon, Drummond 
(1979) found that open-mouthed searching (i.e., 
lateral sweeps with open jaws usually while the 
snake was moving) was used when N. sipedon were 
under water, and this behavior did not require 
visual stimulus, being more likely to occur after 
an unsuccessful attack. These observations of open- 
mouthed searching correspond closely with those 
for a C. rynchops which was seen behaving in this 
fashion at night, in muddy water, and after an 
unsuccessful attack. 

Drummond (1985) isolated visual and mechan- 
ical stimuli for predatory behavior of natricines 
and found that, in the presence of diffuse fish odor, 
visual stimulus could elicit an attack. The role of 
visual stimulus for predation by C. rynchops re- 
mains unclear. Compared to N. sipedon, the eyes 
of C. rynchops appear substantially smaller. The 
C. rynchops that were fed Periophthalmus in Ma- 
laysia only attacked fish that were moving, but 
mechanical and chemical stimuli were also present 
in these trials. Cerberus rynchops that were main- 
tained in the United States for a longer duration 
would attack nonmoving fish. During the daytime, 
some attempts were made to capture C. rynchops 
by reaching down from the seawall. The C. ryn- 
chops were very adept at evading this method of 
capture, and they usually dove below the surface 
of the water even before the hand entered the water. 
Hence, it is clear that C. rynchops can respond to 
visual stimulus within about 1 m. Yet, the fact 
that C. rynchops would attack vibrations caused 
by waterdrops suggests visual stimulus may be 
minimally important for the predatory behavior 
of this species. Future, more controlled studies 
comparing homalopsines, natricines, and hydro- 
phiids should clarify different roles of various 
stimuli on their predatory behavior. 



Growth 

The average growth rate of 0. 165 g/day for this 
small sample of C. rynchops is about one-third the 
estimated rate of 0.49 g/day for the sea snake En- 
hydrina schistosa in this same estuary (Voris, 1 985). 
The growth in sv of 0.42 mm/day for this sample 
is also substantially less than the 1 .0 mm/day es- 
timated for E. schistosa in the first year of life 
(Voris & Jayne, 1979). One potential factor af- 
fecting growth rate is the amount of prey con- 
sumed. The total estimated biomass of prey taken 
by C. rynchops was 514.2 g, which was 6.21% of 



the total biomass (8,282 g) of the snakes that con- 
sumed them. Only 29.6% of the C. rynchops ex- 
amined had stomach contents. Assuming the sam- 
ple of snakes with stomach contents was a random 
subsample of all the snakes collected, one can es- 
timate the biomass (in grams) of all the examined 
snakes by the formula: 8,282 x (100/29.6) = 
27,979. Hence, the corrected ratio of biomass of 
prey consumed to biomass of predator equals 
1.84%. Similar estimates of these ratios can be 
calculated for the data set of 1 04 catfish (Voris & 
Moffet, 1981) consumed by E. schistosa at Muar. 
Enhydrina schistosa consumed an estimated 1 , 1 74 
g of fish which was 9.27% of their total biomass 
of 12,672 g. However, only 19.6% of the E. schis- 
tosa had stomach contents. After correcting for 
percentage of stomach contents, the ratio of total 
prey biomass/predator biomass becomes 1.81% 
for E. schistosa, and this figure is remarkably sim- 
ilar to that of C rynchops. 

These gross estimates of prey consumption ig- 
nore the cost of capturing prey. Cerberus rynchops 
was often sighted actively foraging, and on the 
average it was taking relatively more and smaller 
prey items than E. schistosa. Thus, C. rynchops 
may be a more active forager than E. schistosa. 



Reproduction and Population Numbers 

The lack of a comprehensive collection pro- 
hibits definitive conclusions about the reproduc- 
tive cycle of Cerberus rynchops at Muar. Snakes 
were only preserved sporadically when stomach 
contents were not regurgitated. Two gravid fe- 
males with barely visible embryos were collected 
2-4 December 1985. One female (sv = 67 cm, Ms 
without embryos = 208 g) contained 27 embryos, 
and the combined mass of these eggs was 39 g. 
The other snake (sv = 55 cm, Ms = 127 g) con- 
tained 12 embryos which totaled 20 g. From 1-8 
February 1986, three large females were pre- 
served. Two of them (sv = 62.5, 64 cm) had neither 
embryos nor enlarged follicles. The third female 
(sv = 62 cm, Ms = 163 g) contained 1 8 very early 
embryos weighing 29 g. Hence, the condition of 
these reproductive tracts further supports a sup- 
position of no strong seasonality of reproduction 
for the C rynchops at Muar. 

In contrast to the population at Muar, Saint 
Girons ( 1 972) suggested that the reproductive cycle 
of C. rynchops in Cambodia conformed to that of 
other Cambodian homalopsine species. These 
homalopsines generally start vitellogenesis in No- 



JAYNE ET AL.: CERBERUS RYNCHOPS 



13 



vember, mating probably occurs in December to 
early January, and parturition occurs in May (Saint 
Girons, 1972). For C. rynchops in Java, Bergman 
(1955) found females with eggs in the oviducts in 
March, April, May, July, and October; however, 
some months were not sampled. Smith ( 1 943) re- 
ported sv of newborn snakes ranging from 17.5— 
20.0 cm and brood size ranging from 8 to 26. 
Considering this size of newborn snakes and the 
continual occurrence of snakes between 25 and 30 
cm, it is puzzling that no snakes shorter than 25 
cm were collected. Perhaps births were occurring 
in a different habitat, or there is some very weak 
seasonality of reproduction. 

Enhydrina schistosa, the most common hyro- 
phiid occurring in the Muar estuary, shows marked 
seasonality in reproduction. Voris and Jayne (1 979) 
found that vitellogenesis in this species occurs dur- 
ing November to December, ovulation probably 
occurs in December, and young are born from 
mid-February through March. Hydrophis melan- 
osoma, H. brookii, and H. torquatus are the next 
most common hydrophiids at Muar, and their re- 
productive cycle is similar to that of E. schistosa 
(Lemen & Voris, 1981). Limited data are available 
for the reproductive cycle of Acrochordus granu- 
latus at Muar. However, collections of A. granu- 
latus from two sites on the west coast of Malaysia, 
one within about 24 1 km of Muar and the other 
80 km from Muar, suggest this species is aseason- 
ally reproductive (Voris & Glodek, 1980). 



Acknowledgments 

We wish to thank the Department of Biology of 
the University of Malaysia for its help. Dr. E. O. 
Murdy kindly assisted in the field, as well as as- 
sisting with identification of fish species. We also 
thank Carole Jayne for her enthusiastic assistance 
in the fieldwork and Helen Voris for her editorial 
comments. Clara Richardson skillfully prepared 
the figures. Financial support for this research came 
from a gift from the Allen-Heath Memorial Foun- 
dation and a grant (no. INT-8305817) from the 
National Science Foundation. 



Literature Cited 

Bailey, N. T. J. 1952. Improvements in the interpre- 
tation of recapture data. Journal of Animal Ecology, 
21: 120-127. 



Bergman, R. A. M. 1955. L'anatomie de Cerberus 
rynchops. Archives Neerlandaises de Zoologie, 11:113- 
126. 

Drummond, H. M. 1979. Stimulus control of am- 
phibious predation in the northern water snake (Ner- 
odia s. sipedori). Zeitschrift fur Tierpyschologie, 50: 
18^4. 

. 1985. The role of vision in the predatory be- 



haviour of natricine snakes. Animal Behavior, 33: 206- 
215. 

Dunson, W. A., ed. 1975. Biology of the Sea Snakes. 
University Park Press, Baltimore, 530 pp. 

Dunson, W. A., and M. K. Dunson. 1979. A possible 
new salt gland in a marine homolopsid snake {Cerberus 
rynchops). Copeia, 1979(4): 661-672. 

Dunson, W. A., and S. A. Minton. 1978. Diversity, 
distribution, and ecology of Philippine marine snakes 
(Reptilia, Serpentes). Journal of Herpetology, 12(3): 
281-286. 

Glodek, G. S., and Voris, H. K. 1982. Marine snake 
diets: Prey composition diversity and overlap. Copeia, 
1982(3): 661-666. 

Gyi, K. K. 1970. A revision of colubrid snakes of the 
subfamily Homalopsinae. University of Kansas Pub- 
lications of the Museum of Natural History, 20(2): 47- 
223. 

Heatwole, H., and R. S. Seymour. 1978. Cutaneous 
oxygen uptake in three groups of aquatic snakes. Aus- 
tralian Journal of Zoology, 26: 48 1—486. 

Jayne, B. C. 1985. Swimming in constricting (Elaphe 
g. guttatta) and nonconstricting (Nerodia fasciata pic- 
tiventris) colubrid snakes. Copeia, 1985(1): 195-208. 

Kardong, K. V. 1980. Evolutionary patterns in ad- 
vanced snakes. American Zoologist, 20: 269-282. 

Kropach, C. 1975. The yellow-bellied sea snake in the 
eastern Pacific, pp. 186-213. In Dunson, W. A., ed., 
Biology of the Sea Snakes. University Park Press, Bal- 
timore. 

Lemen, C. A., and H. K. Voris. 1981. A comparison 
of reproductive strategies among marine snakes. Jour- 
nal of Animal Ecology, 50: 89-101. 

McCosker, J. E. 1975. Feeding behavior of Indo- Aus- 
tralian Hydrophiidae, pp. 217-232. In Dunson, W. 
A., ed., Biology of the Sea Snakes. University Park 
Press, Baltimore. 

Pickwell, G. V. 1 972. The venomous sea snakes. Fau- 
na (Rancho Mirage, Calif.), no. 4: 17-32. 

Pough, F. H., and J. D. Groves. 1 983. Specializations 
of the body form and food habits of snakes. American 
Zoologist, 23: 443^154. 

Saint Girons, H. 1972. Les serpents du Cambodge. 
Memoires du Museum National d'Histoire Naturelle, 
Nouvelle serie, Serie A, Zoologie, 74: 1-170. 

Schoener, T. W. 1968. The Anolis lizards of Bimini: 
Resource partitioning in a complex fauna. Ecology, 
49: 704-726. 

Smith, M. A. 1943. The Fauna of British India. Taylor 
and Francis, London, 583 pp. 

Tweedie, M. W. F. 1983. The Snakes of Malaya, 3rd 
ed. Singapore National Printers Ltd., Singapore, 167 
pp. 



14 



F1ELDIANA: ZOOLOGY 



Voris, H. K. 1985. Population size estimates for a sea 
snake (Enhydrina schistosa) in Malaysia. Copeia, 
1985(4): 955-961. 

Voris, H. K., and G. S. Glodek. 1980. Habitat, diet, 
and reproduction of the file snake, Acrochordus gran- 
lulatus in the straits of Malacca. Journal of Herpetol- 
ogy, 14(1): 108-111. 

Voris H. K., and B. C. Jayne. 1979. Growth, repro- 
duction and population structure of a marine snake, 
Enhydrina schistosa (Hydrophiidae). Copeia, 1979(2): 
307-318. 

Voris, H. K., and M. W. Moffett. 1981. Size and 



proportion relationships between the beaked sea snake 
and its prey. Biotropica, 13(1): 15-19. 

Voris, H. K.., and H. H. Voris. 1983. Feeding strategies 
in marine snakes: An analysis of evolutionary, mor- 
phological, behavioral and ecological relationships. 
American Zoologist, 23: 41 1-425. 

Voris, H. K., H. H. Voris, and Lim Boo Liat. 1978. 
The food and feeding behavior of a marine snake, 
Enhydrina schistosa. Copeia, 1978(1): 134-146. 

Wall, F. 1918. A popular treatise on the common 
Indian snakes. Journal of the Bombay Natural History 
Society, 26: 89-97. 



JAYNE ET AL.: CERBERUS RYNCHOPS 



15 




Field Museum of Natural History 
Roosevelt Road at Lake Shore Drive 
Chicago, Illinois 60605-2496 
Telephone: (312) 922-9410 



UNIVERSITY OF ILLINOIS URBANA 

590.5FIN.S. C001 

FIELDIANA : ZOOLOGY $ NEW SERIES $CHG0 

40-54 1988-89 




3 0112 009378735